A Year of Transformation: From an Empty CRM to an AI-Powered Search Operation
One of CB Consulting's longest-standing partnerships is with a boutique, founder-led executive search firm in the US. We started working together just over a year ago, and in that time the engagement has grown from a single, focused project into an ongoing, evolving partnership — the kind of relationship this business is built to support.
Twelve months on felt like the right moment to look back at what "transformation" actually looks like in practice. Not as a slide of buzzwords, but as a sequence of real decisions, trade-offs, work and real results.
The Challenge
Like a lot of specialist search firms, my client's needs were growing faster than its systems. Candidate and client records lived in a legacy CRM/ATS that had never been given a proper clean-up — years of history, no consistent structure, and little compliance-ready hygiene. Every candidate update, every search report, and every piece of research was still being handled manually by a lean team that was already stretched thin.
At the same time, AI entered the room in a very real way. With untold potential possible, it was clear from the start we had to build an AI strategy that was both intentional and flexible. It needed to match the pace of the business’ needs, yet also leave room for exploration and testing to blend with the rapidly changing tech landscape. With a few initial tools purchased, we set out to see what we could build.
The Approach
While this initiative began as a single fixed-scope project, we soon discovered that designing a rolling partnership was the next natural step — deliberately building on each phase completed made the next one possible.
Phase one was foundation. Before anything clever could happen, the CRM/ATS needed to actually work. That meant re-designing candidate and job-stage workflows and compliance structures from scratch, migrating and cleaning years of historic data from multiple old systems (lots of excels!), and running proper onboarding so the change would stick — followed by a recurring health-check cadence to keep it that way.
Phase two was automation. Once the data could be trusted, we built an end-to-end reporting pipeline that pulls live CRM/ATS data and turns it into finished report documents — delivered automatically, with minimal manual adjusting, but grounded in human-led approval. Alongside it came AI-drafted candidate update emails, AI-generated candidate summaries with safeguards against inaccurate output, and a reorganised, permissioned delivery structure so that each clients retained access securely.
Phase three was AI-empowered systems. Connecting their tool suite via MCP with their preferred LLM, Claude, was the next natural (and timely!) step. Getting the tool suite to talk to each other provides that refinement layer on top of a solid (now organised) knowledge base, and unlocks continuous additional value and insights.
Phase four will be focused on iteration and scale. With the foundation solid, new workflows being built, and a refinement layer across the top, we have overhauled the way my client does search. To building a stronger research capability feeding into new searches and creating new insights for clients, to trusting their data for actual insights on search progress, to demonstrating expertise and human-led analysis, it's exciting to consider what else we will unlock along the way.
Adjusting Throughout
Priorities shift, a search goes awry, clients’ needs change, a new update/ functionality is rolled out; it's safe to say we've seen a few adjustments along the way. One of the benefits of a rolling engagement means we are able to course correct without dismantling the overall mission or goal. The foundation and progress made remains; but it's the flexibility that allows us to adjust focus where and when needed.
The Results
A year in, the shift is tangible:
One clean system of record, replacing a fragmented legacy setup that no one fully trusted.
How it's mattered > consistency and trust in data = less time looking for scattered information and producing a successful search outcome.
Multiple automated reporting pipelines, producing polished, client-ready output with none of the manual drafting that used to eat hours every week.
How it's mattered > ~20 mins is now the average report build time (versus ~2 hours!). One of which was so good, their client insisted all future reporting from all other talent partners use the same format (!).
AI-assisted communication, cutting the time spent writing candidate updates and summaries without sacrificing quality or accuracy
How it's mattered > saves ~1-2 hours of time per day AND speeds up client analysis of relevant candidate profiles.
A renewed research process, backed by a solid human-led understanding of “what good looks like” — so institutional knowledge lives in a system, not just in someone's head.
How it's mattered > when a search recently pivoted, 2 previously-undiscovered, viable and interested candidates were screened within 3 days of the pivot - with more waiting in the pipeline.
A connected technology stack, where CRM, automation, and research tools actually talk to each other instead of operating as five separate silos.
How it's mattered > previously undiscovered insights are now at their finger tips…and a wide variety of additional automations are now being tested and implemented!
What made the difference?
None of this happened in one sprint. It happened purposefully, and in the right order — foundation before testing & automation, automation before iteration and scale — with enough trust built along the way that the work could keep evolving rather than stopping at "project complete."
That's really the point of this piece. AI transformation in executive search isn't a single tool you buy or a weekend of training, or even 3 months of “trying” new tools. It's a sequence of unglamorous, deliberate decisions, sustained over time, with clients willing to invest in getting the basics right before chasing what's popular. The ROI lies within the nuggets of value you unearth along the way, all of which feed into every element of the process.
If your firm's knowledge assets feel more like a collection of disconnected tools than a genuine system, and you're wondering where to start, or how to think about timelines; let's have a conversation. It's always harder to view a goal from the endpoint working backward; so let's take it step by step.
